PT - JOURNAL ARTICLE AU - Samaneh Kouchaki AU - Avraam Tapinos AU - David L Robertson TI - Metagenomic Binning through Multi-resolution Genomic Binary Patterns AID - 10.1101/096719 DP - 2016 Jan 01 TA - bioRxiv PG - 096719 4099 - http://biorxiv.org/content/early/2016/12/24/096719.short 4100 - http://biorxiv.org/content/early/2016/12/24/096719.full AB - Motivation High-throughput sequencing has facilitated the analysis of complex microbial communities. Consequently, an enormous number of sequences have been generated containing various regions of bacterial and viral genomes. Image processing offers a rich source of descriptors for data analysis. Here, we introduce a feature space called multi-resolution local binary patterns (MLBP) from image processing as a feature descriptor to extract local ‘texture’ changes from nucleotide sequences. We demonstrate its applicability to the alignmentfree binning of metagenomic data.Results The effectiveness of our approach is tested using both simulated and real human gut microbial communities. We compared the performance of our method with several existing techniques that are based on k-mer frequency to show it outperforms existing techniques. In addition, we provide a time-series study of the abundance pattern of each bin to help refine the formed clusters automatically and to find relations that may exist among the clusters. Although the main aim is to introduce the use of genomic signatures using an alternative feature space (MLBP), our results show its application to the analysis of contigs from a metagenomic study.Availability The source code for our Multi-resolution Genomic Binary Patterns method can be found at https://github.com/skouchaki/MrGBP